AI systems are being built and deployed at an unprecedented pace, yet there's no single home for the people building them to come together, share discoveries, and establish best practices. The work spans architecture research, production infrastructure, and agent design, scattered across venues that each cover a slice. CAIS is our attempt to put it all under one roof.

The demo track is at the heart of this. We want to see real systems doing real things. Show us the agent that coordinates across tools and APIs in ways we haven't seen before. Show us the serving infrastructure that makes compound AI systems practical at scale. Show us the debugging tool that finally makes production AI observable. Whether it's a research prototype or a production system, if it pushes the field forward, we want it here.

Topics of Interest

We invite demonstrations across two broad categories, though we encourage submissions that blur the lines between them:

Agent Demonstrations

• Autonomous AI agents with formalized goal-driven behavior

• Novel multi-agent systems with emergent coordination and interaction protocols

• Integration of agents with external real-world systems, sensors, or hardware

• Evaluated agent behavior with principled benchmarking or user studies

• Agents with persistent memory, learning, and adaptation over time

• ...and more! If your agent does something we haven't seen before, we want to hear about it.

AI Systems Demonstrations

• Novel inference serving systems with efficiency or scalability advances

• Compound AI system architectures with principled composition

• Data and ML pipeline systems for training, fine-tuning, or curation at scale

• Production AI system monitoring, evaluation, and debugging tools

• Optimization and adaptation of AI systems for deployment constraints

• ...and more! If you've built infrastructure that makes AI systems faster, more reliable, or more practical to operate, show us.

These categories are starting points, not boundaries. Some of the most interesting work sits at the intersection of agents and systems. If you've built something that crosses both, or doesn't fit neatly into either, we still want to see it.

The submission details for the CAIS 2026 System Demonstrations track are now available. For those of you who have been asking, here's what you need to know:

Submission format:

• 4-page paper (+ unlimited references/appendices) in ACM sigconf format, single-blind

• 3-5 minute demo video required

• Live demo artifact strongly encouraged

We welcome demos across agent systems (autonomous agents, multi-agent coordination, real-world integration, persistent memory) and AI systems (inference serving, compound AI architectures, ML pipelines, monitoring tools, deployment optimization), and beyond. If you've built something novel at the intersection of AI and systems, we want to see it.

Demos can emphasize technical depth, compelling interactive experience, or both. Reviewers will evaluate each submission based on its strengths.

Key dates:

• Demo submission deadline: Friday, March 13, 2026 (11:59 PM AoE)

• Notification of acceptance: Friday, April 10, 2026

• Camera-ready deadline: Monday, April 27, 2026 (11:59 PM AoE)

A note on submissions: We are aware that some demo abstracts were registered on the HotCRP site (acm-cais26.hotcrp.com), which is intended for research papers only. If you submitted a demo there, please email program-chairs@caisconf.org with your paper number, title, and author list so we can locate your submission. We will follow up with details on the correct submission process for demos.

Full details including submission guidelines and review criteria:
https://caisconf.org/pages/demos/

Best regards,
Heather Miller, Avi Sil, and Omar Khattab
CAIS'26 Program Chairs